NAES Data for 2004

I need to take a moment to note something I overlooked during
the very busy final days of last fall's campaign. In late October, the National Annenberg Election
Survey (NAES) released its complete respondent level data collected during
the 2004 campaign, which can now be purchased for $35 on CD-ROM along with a textbook on how to analyze the data,
Capturing Campaign
Dynamics 2000 & 2004.

For those unfamiliar, NAES is an academic survey project
that conducted a nightly telephone tracking survey throughout both the 2000 and
2004 campaigns. The 2004 tracking started in October 2003 and continued nightly
through mid-November 2004, completing 81,422 interviews as part of their
national "rolling cross sectional" design. The details are explained on their methodology
page, but the gist is that they applied the most rigorous methods of
telephone survey research, and the net result is (as the book jacket claims) "the
largest studies ever undertaken of the American electorate. It averaged more
than 160 interviews per day (about 5,000 per month) during the primary season,
and 300 per day (nearly 9,000 a month) during the final two months of the
campaign.

The 2004 project also includes additional data sets: NAES interviewed
special "over-samples" among voters in New
Hampshire just before the primary and in the households
of in active duty military personnel in the fall. They also conducted panel
studies (which interview the same individuals before and after an event) to
track changing opinions during the political conventions, the debates and the
general election itself.

And to be clear, the CD that comes with the new book includes
all of the data from both the 2000 and 2004 surveys, over 200,000 interviews in
all. As a tool for academics and political junkies that know their way around
statistical software, there is nothing else quite like it.